Engineers who turn foundation models into products people use. Less model training, more retrieval, evaluation, tooling and the software around the model.
Applied AI Engineers build features on top of large language models and other foundation models. The work sits between software engineering and machine learning: designing prompts and retrieval pipelines, wiring models into existing systems, and measuring whether the output is good enough to ship.
Most UK openings are at AI-native scaleups, product companies adding AI features to existing software, and the AI teams inside banks, insurers and consultancies. Employers usually want a strong software engineer first; deep research experience is rarely required, but a working grasp of how models fail is.
Most weeks are product engineering: building and testing features, with a steady share of time spent on evaluation and on the failures users report.
Based on advertised midpoints across the 19 UK listings priced in pounds in the last 12 months. Base salary only.
% of 30 listings posted in the last 12 months that mention each skill, extracted from job descriptions.
A typical UK progression for applied ai engineers. Years are guidance — strong people move faster, and many senior folks sidestep into research, product or management.
Builds and tests individual features on an existing model stack, with designs reviewed by senior engineers.
Owns a product area end to end, from retrieval design to evaluation and production monitoring.
Sets the technical approach for a team, chooses models and infrastructure, and owns quality and cost targets.
Shapes how the organisation builds with AI: shared platforms, evaluation standards and model strategy.
There's no single route, but most people follow some version of these steps.
Most applied AI work is software engineering. Python or TypeScript services, APIs and data handling come first.
Understand tokens, context limits, embeddings and the common failure modes of generation.
Ship a retrieval or agent project with a real evaluation set, not just a demo.
Join a team shipping AI features to users, where evaluation and cost matter.
Lead evaluation, monitoring and model choice for a product used in production.
ElevenLabs
Builds agents that run real parts of the business, working in Marketing Engineering and forward-deployed to growth pods to find problems and own solutions end to end. Uses Python, SQL, LLM APIs and agent orchestration, with a focus on making agents reliable, measurable and safe enough for the business to depend on.
LLM APIs
Anthropic · 51–200 people · Series B
Join the Beneficial Deployments Life Sciences team to help labs automate experiments end-to-end with Claude, connecting software to robots, instruments and LIMS. You'll prototype and productionise LLM-powered tools, build evals, contribute to the Model Hardware Standard, and scale learnings through cookbooks and docs. The role involves significant travel and acting as an EMEA ambassador for Claude in scientific settings.
Claude · Model Hardware Standard · Liquid Handlers +2
OpenAI
Partner directly with Digital Native customers to design, build and deploy AI-powered products, taking ideas from early experimentation through evaluation, production launch and scale. Hands-on coding in Python to build prototypes, evaluation harnesses, integrations and production accelerators, while making decisions across models, agents, retrieval, reliability, latency, cost, safety and governance.
OpenAI
Anthropic · 51–200 people · Series B
You'll act as a trusted technical advisor to high-growth startups building on the Claude Developer Platform, pairing with customer engineering teams on architecture reviews, code contributions and agent design. The role involves creating evaluation frameworks, prototypes and developer content, plus travel to customer sites for workshops and implementation support.
Claude · Claude Developer Platform
OpenAI
You'll partner with early-stage startups to embed Codex into their software workflows, building integrations, automations and reusable tools while troubleshooting technical blockers. A large part of the role is outward-facing: leading technical talks, live demos, workshops and hackathons, then feeding developer insights back to OpenAI's product and research teams.
Codex · OpenAI
Scale AI
Sets standards for responsible AI, model governance and production MLOps across a public sector team, and builds or validates the highest-risk parts of strategic AI systems. Acts as senior technical partner to client leadership, coaches engineers and leads resolution of severe model safety or data integrity incidents. Requires deep experience with regulated, sovereign or on-premise deployment and auditability.
Scale AI
Own the AI behaviour of production systems for public sector clients, covering agents, model integration, evaluation design and the backend services they run on. Lead delivery for a workstream, act as the technical client contact on live-data performance, and mentor other engineers.
Deerfoot Recruitment Solutions
Architect and build multi-agent AI systems end-to-end, combining frameworks like LangGraph or Haystack with bespoke orchestration, state management and tool-calling logic. Embed LLMs and vision-language models into live reasoning, search and task-execution workflows, deploying across cloud and on-premises environments with monitoring, evaluation and guardrail tooling. Set technical standards for agentic AI within the team, working with a biotech and pharma customer.
LangGraph · LangChain · Haystack +7
The locations with the most live listings for this role today.
Usually not. Employers mostly hire strong software engineers who can show working projects built on models, with sensible evaluation.
Machine Learning Engineers more often train and serve their own models. Applied AI Engineers mainly build products on existing foundation models.
Python is the most common, with TypeScript close behind for product teams. SQL and cloud experience appear in most adverts.
Pay varies with seniority, sector and location. The salary section on this page is drawn from live adverts on the board.
Post your role in 90 seconds and reach the specialist audience that already reads this page.
